Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site arizona.UUCP
Path: utzoo!linus!security!genrad!grkermit!masscomp!clyde!floyd!harpo!ihnp4!arizona!budd
From: budd@arizona.UUCP
Newsgroups: net.lang.st80
Subject: ambiguous grammar
Message-ID: <7467@arizona.UUCP>
Date: Fri, 20-Jan-84 18:01:18 EST
Article-I.D.: arizona.7467
Posted: Fri Jan 20 18:01:18 1984
Date-Received: Sat, 21-Jan-84 07:34:40 EST
Organization: CS Dept, U of Arizona, Tucson
Lines: 20

on the ambiguous grammar -

no, you miss my point.  Look, for example, at page 79 in the Goldberg book:

   (incomes includesKey: source)
        ifTrue: [...]
        ifFalse: [ ^ 0]

   totalSpentFor: reason

now, the totalSpentFor is the start of the NEXT method.  However, looking
at the printed source, and attempting to parse it, there is no way that
one can tell this from the possibility that totalSpentFor is part of the
command being sent to (incomes ...).

The answer seems to be that the Smalltalk system does NOT look at the source
as it is displayed, but rather keeps it internally in a different form, and
edits, etc are performed at the method level, not at the entire class
description level. nmk@brown has pointed this out to me by mail, but
subsequent messages on the net seem to have missed the point.